Epidemiological studies show that preexisting insomnia increases the odds of incident anxiety disorders and vice versa. We have suggested that preexisting sleep disturbance can disrupt sleep-dependent emotion regulatory circuits, such as those supporting consolidation of fear extinction memory, leading, in turn, to greater daytime anxiety symptoms. Exacerbation of both sleep disruption and anxiety symptoms can then occur via a positive feedback (?vicious cycle?) mechanism. The proposed research will test this 3- component mutual-reinforcement model in Generalized Anxiety Disorder (GAD), the anxiety disorder most commonly comorbid with Insomnia Disorder (ID). In a sample of individuals with GAD having varying degrees of sleep disturbance, we will measure associations between each pair of the above elements ? 1) sleep quality, 2) extinction memory and its associated neural substrates and 3) GAD symptomatology ? to determine whether such associations support this positive feedback model. In a recent fMRI study comparing individuals with Primary Insomnia (PI) to age and sex-matched good sleeping healthy controls (HC) and individuals with GAD, we have identified two factors that might predispose persons with ID to develop GAD. First, using resting state functional connectivity, we found that individuals with PI showed connectivity between the left amygdala and the rostral anterior cingulate cortex (a region believed to regulate amygdala activity) that was both significantly less than HC and intermediate between that of HC and individuals with GAD. Thus emotion regulation may be weakened in PI thereby increasing vulnerability to GAD. Second, using a validated 2-day fear conditioning and extinction protocol, we identified delayed engagement of extinction-related structures in PI relative to HC. In HC, activation of extinction-related structures (ventromedial prefrontal cortex and hippocampus) occurred during extinction learning, but did not do so in PI until extinction recall 24 h later. Research will examine 50 individuals with GAD half of whom can be diagnosed with comorbid ID and half of whom will not meet criteria for ID. Following psychiatric and sleep-disorders screening interviews, participants will complete 14 days of longitudinal actigraph and sleep-diary monitoring, an online battery of anxiety and sleep quality assessments, three nights of ambulatory polysomnography and two fMRI scanning sessions. fMRI will consist of both resting state scans and the above fear conditioning and extinction protocol with simultaneous measurement of skin conductance. The three components of the positive feedback model will be analyzed from both a categorical (GAD with vs. without co-morbid ID) and a dimensional perspective that accords with the NIMH Research Domain Criteria approach. General hypotheses tested will be 1) that co-morbid insomnia in GAD worsens behavioral and neural aspects of fear extinction as well as resting state-connectivity between fear expression and emotion regulatory structures, and 2) that poor sleep quality, poor extinction memory and greater GAD symptomatology show the predicted mutual exacerbation predicted by the positive feedback model.